© 2000 American Heart Association, Inc.
ATVB Electronic Pages |
Fibronectin in an Extracellular Matrix of Cultured Endothelial Cells Supports Platelet Adhesion via Its Ninth Type III Repeat
A Comparison With Platelet Adhesion to Isolated Fibronectin
From University Hospital Utrecht, Department of Hematology, Utrecht, Netherlands.
Correspondence to Dr J.J. Sixma, University Hospital Utrecht, Department of Hematology (G03.647), Heidelberglaan 100/PO Box 85500, 3584 CX Utrecht/3508 GA Utrecht, Netherlands. E-mail jsixma{at}laboratory.azu.nl
Abstract
Abstract—We investigated the involvement of different domains of fibronectin in mediating platelet adhesion to fibronectin in the extracellular matrix (ECM) of cultured endothelial cells under flow conditions. Polyclonal anti-fibronectin antibodies were absorbed with Sepharose to which no protein, intact fibronectin, or different fibronectin fragments had been coupled to obtain supernatants (Sups) (Sup0, SupFN, and Supname of the fragment, respectively) from which a specific part of the antibodies had been removed. Treatment of the ECM before perfusion with Sup0 resulted in a 36% decrease in platelet coverage, whereas treatment with SupFN resulted in maximal adhesion. Treatment of the ECM with supernatants from which antibodies directed against the gelatin- or heparin-binding domain had been removed showed the same inhibition as treatment with Sup0. Removal of antibodies directed to the 120-kDa cell-binding domain resulted in a level of adhesion equal to the level found when the ECM was treated with SupFN. Further analysis of this central region showed that only treatment with supernatants from which antibodies directed to the ninth type III repeat (III-9) of fibronectin had been removed resulted in a significantly higher adhesion than treatment with Sup0. Studies of adhesion to the fragments themselves showed that only fragments containing III-10 were able to support adhesion. Mutation of the Arg-Gly-Asp (RGD) sequence into Arg-Gly-Glu (RGE) in one of those fragments resulted in a complete loss of adhesive capacity. These data suggest that platelet adhesion to fibronectin in the ECM depends on III-9, whereas III-10 does not seem to be required. For platelet adhesion to isolated fibronectin, an intact RGD sequence seems to be crucial.
Key Words: platelet adhesion • fibronectin • endothelial cell matrix
The extracellular matrix (ECM) protein fibronectin is involved in a variety of biological processes by mediating cell adhesion and migration.1 As a constituent of the subendothelium of the vessel wall, it is recognized by blood platelets. In this way, fibronectin contributes to the process of hemostasis, which follows after the vessel has been damaged and the integrity of the endothelial cell layer has been lost.
Fibronectin is composed of 3 types of homologous repeats, designated as
types I, II, and III.2 Proteolysis yields
protease-resistant functional domains that interact with
heparin, collagen, fibrin, and cells. The cell-binding domain, which
occupies the central region in the molecule, consists of type III
repeats, each 
90 amino acids in length.3
The first sequence in fibronectin found to possess cell-adhesive
properties was the arginine-glycine–aspartic acid (RGD) sequence,
which is located in the 10th type III repeat (III-10) of the
cell-binding domain.4 On platelets, this sequence in
fibronectin is recognized by 2 receptors, glycoprotein (GP)
IIb/IIIa5 6 and very late antigen 5 (VLA-5),7
corresponding to GP Ic/IIa on the platelet.8 9 10 Both
receptors are members of the superfamily of integrins, a group of cell
surface receptors composed of noncovalently associated 
- and
ß-subunits.11 However, several studies have suggested
that additional sequences are needed for optimal cell-adhesive
activity. Proteolytic fragments of the cell-binding domain >75 kDa
show adhesive activity equal to that of intact fibronectin, whereas an
11.5-kDa fragment and smaller synthetic peptides display a 20- to
100-fold loss of activity relative to intact
fibronectin.4 12 13 14 Information on the nature of these
additional sites has become available from different sources. By use of
deletion mutants of fibronectin or specific antibodies,
VLA-5–dependent cell spreading has been described to depend on sites
in III-7 and/or III-8,15 III-8 and/or
III-9,16 and III-8 and III-9.17 18 19 For GP
IIb/IIIa, additional sites have been mapped to III-920 21
and to the amino-terminal part of III-10.20
In a previous study, we found that platelet adhesion to isolated surface-immobilized fibronectin under flow conditions was partially inhibited by antibodies directed to VLA-5, whereas an antibody directed to GP IIb/IIIa and an RGD-containing peptide inhibited adhesion almost completely.22 In contrast, platelet adhesion to a matrix of cultured endothelial cells (ECM) at a low shear rate of 300 s-1 was not inhibited by these antibodies and the peptide, although adhesion partially depended on fibronectin present in the ECM.22 23 In this study, we investigated the role of different fibronectin domains in mediating platelet adhesion to fibronectin in an ECM, and we compare this with adhesion to the isolated protein. We show that fibronectin-dependent adhesion to the ECM involves III-9 of the central cell-binding domain, whereas III-10 does not seem to be required. For adhesion to the isolated protein, an intact RGD sequence seems to be crucial.
Methods
Fibronectin Fragments
Proteolytic Fragments of Fibronectin
Purified human fibronectin was prepared from citrated plasma as
described previously by affinity chromatography over a
gelatin-Sepharose (Pharmacia) column.24 Proteolytic
fragments of fibronectin were obtained by digestion with cathepsin D as
previously described,25 with some modifications. Briefly,
fibronectin in a 50 mmol/L sodium acetate buffer (pH 3.5)
containing 50 mmol/L NaCl, 10 mmol/L EDTA, 0.2 mmol/L
PMSF, and 0.02% sodium azide was digested for 3 hours at 37°C with
cathepsin D (Sigma Chemical Co; 1:300, wt/wt, enzyme/substrate). The
reaction was stopped by raising the pH to 7.5 with a 2.5 mol/L Tris/HCl
solution, pH 8.5, and the conductance of the digest was reduced to 6
mho by addition of distilled water. Then the digest was applied to a
gelatin-Sepharose column, and the flow-through of this column was
passed over a heparin-Sepharose (Pharmacia) column. Both columns were
equilibrated with buffer A (20 mmol/L Tris [pH 7.5] containing
50 mmol/L NaCl, 10 mmol/L EDTA, 0.2 mmol/L PMSF,
and 0.02% azide). After extensive washing, the gelatin-binding domain
was eluted from the gelatin-Sepharose with 50 mmol/L Tris
containing 6 mol/L urea and 100 mmol/L citric acid (pH 4.7).
Fractions were pooled, dialyzed against 50 mmol/L Tris/100
mmol/L NaCl (pH 7.5), and stored at -20°C. The cell-binding domain
and the high-affinity heparin-binding domain were eluted from the
heparin-Sepharose with buffer A containing 0.1 and 0.5 mol/L NaCl,
respectively. Fractions were pooled and stored at -20°C.
Fusion Proteins
Fusion proteins with glutathione-S-transferase (GST)
were constructed in the expression vector pRP265, a derivative from the
PGEX-2T vector,26 in which the following polylinker
has been cloned into the BamHI/EcoRI site of the
original vector:
GGATCCCCATGGTACCCGGGTCGACTAGTATGCATAAGCTTGAATTC
BamHI Kpnl AccI NsiI
HindIII EcoRI (gift from Dr C. Vink and Dr R.H.A.
Plasterk, Netherlands Cancer Institute, Amsterdam). All fragments were
produced by polymerase chain reaction (PCR). Plasmid pFH100 (EDII-,
EDI-, EDIIICS89), originally constructed by Dufour et
al27 and provided to us by Dr D.F. Mosher (University
of Wisconsin, Madison), served as a template. Table 1
shows the primers that were used
in the reactions. Fragment FN6 RGE, in which the RGD sequence has been
changed into RGE, was produced as 2 PCR fragments (fragments A and B)
in which the internal primers cover the site of the mutation (which is
underlined). PCR products were separated on a 1% agarose gel, and
DNA fragments of the right size were extracted from the gel with
Qiagen. PCR products of fragments FN4, FN5, FN6, and FN7 were
digested with restriction enzymes and ligated in the respective sites
in the vector. PCR products of fragments FN, FN2, FN3, and FN6 RGE
were directly inserted into the linearized vector by the exonuclease
recession technique.28
|
Fragments were cloned into the following sites of the vector: FN4 and FN5 into the KpnI/NsiI sites, FN1, FN2, FN3, FN6, and FN6 RGE into the KpnI/HindIII sites, and FN7 into the BamHI/AccI sites. The authenticity of all fragments was confirmed by both restriction analyses and sequencing according to the method of Sanger et al.29 The mutation in FN3 RGE was confirmed by restriction analysis using BanII. Proteins were expressed in Escherichia coli and purified out of the soluble fraction of the bacterial lysate with glutathione–Sepharose 4B according to the instructions of the manufacturer (Pharmacia). GST alone was used as a control in our experiments.
The locations of all fragments, including the recombinant 33-kDa and
the 40-kDa fragments, which were a generous gift from Dr T. Vogel
[BioTechnology General (Israel) Ltd, Rehovot, Israel],30
are depicted in Figure 1.
|
Antibodies
Polyclonal rabbit anti-human plasma fibronectin antibodies
[F(ab')2 fragments] were purchased from Cappel
Organon Teknika Corp. The F(ab')2 fragments were
prepared from the total IgG fraction in the rabbit serum, including IgG
inherent in the rabbit. Peroxidase-conjugated polyclonal swine
anti-rabbit antibodies were obtained from Dakopatts. Monoclonal
antibodies (MAbs) directed to the cell-attachment site or high-affinity
heparin-binding domain in fibronectin were purchased from
Boehringer Mannheim. Polyclonal antibody
anti–G-TM3,31 which recognizes GST, was a gift from Dr A.
de Ronde (Institute of Virology, Faculty of Veterinary Medicine,
Utrecht University, Netherlands).
Protein Characterization
SDS-PAGE was performed according to Laemmli,32 with
the Pharmacia Phastgel system. After purification of the fusion
proteins on glutathione beads, we determined the proportion of intact
protein in the sample with densitometry (Molecular Dynamics) performed
on gels stained with Coomassie brilliant blue. For Western blotting,
proteins were transferred to a PVDF membrane (Immobilon-P) by
electroblotting. After blotting, the membrane was blocked with a 5%
solution of Protifar (Nutricia) in Tris-buffered saline (TBS; 10
mmol/L Tris, 0.9% NaCl, pH 7.4) for 30 minutes. Then the blot was
incubated for 1 hour with 5 µg/mL polyclonal rabbit anti-human
fibronectin antibodies in the same solution. After extensive washing
with TBS containing 0.5% Tween 20, the blot was incubated with
peroxidase-conjugated swine anti-rabbit antibodies, diluted 1000:1 in
the blocking solution. After another extensive washing, the blot was
stained with a substrate solution consisting of 5 mg of
3,3'-diaminobenzidine (Sigma Chemical Co) per 50 mL of 50 mmol/L
Tris (pH 7.4) to which was added 5 µL of 30% hydrogen peroxide and
0.03% NiCl2.
Surfaces
Surface of Intact Fibronectin or Fibronectin Fragment
Glass coverslips (18x18 mm; Menzel) were cleaned in 80%
alcohol, rinsed in distilled water, and dried thoroughly. Per
coverslip, 100 µL of a solution of an indicated concentration of
intact fibronectin or fibronectin fragment, dialyzed against 0.1 mol/L
ammonium acetate (pH 7.4), was sprayed with a retouching airbrush
(Badger model 100, Badger Brush Co) in such a way that each layer of
protein had been dried before the next layer was applied. For the
fragments, we corrected the calculated concentration for the proportion
of intact fragment present in the protein sample as determined by
densitometry. After spraying, the glass coverslips were incubated with
a 1% human albumin solution for 1 hour to block aspecific
adhesion to glass. No adhesion was found on glass coverslips coated
with albumin alone.
Extracellular Matrix
Human umbilical vein endothelial cells were
isolated and grown to confluence as described.33 Cells of
the second passage were seeded on glass coverslips coated with gelatin.
Cells were removed with 0.1 mol/L NH4OH for 15
minutes at room temperature, and subsequently the matrices were washed
3 times with PBS (10 mmol/L phosphate buffer [pH 7.4] and 0.15
mol/L NaCl). The isolation procedure of the ECM did not influence the
amount of fibronectin in the ECM,34 and the matrix is free
of cell membrane fragments.35
Preparation of Antibody Supernatants
Polyclonal rabbit anti-human fibronectin antibodies, which had
previously been shown to completely inhibit adhesion to purified
fibronectin,22 were dissolved in PBS and incubated for 2
hours with intact fibronectin, a fibronectin fragment, or as control
for the fusion proteins, GST, which had been coupled to
CNBr-activated Sepharose 4B according to the instructions of
the manufacturer (Pharmacia). The antibodies were also incubated with
CNBr-activated Sepharose that had gone through the whole
coupling procedure but to which no protein was coupled. Then the
sepharoses were spun down, and the supernatants were collected. Each
supernatant (Sup) was named after the protein it had been in contact
with (Sup0 for incubation with Sepharose to which
no protein was coupled; SupFN,
SupGST, and Supfibronectin
fragment for incubation with intact fibronectin, GST, or a
fibronectin fragment, respectively; Supgelatin,
Supcell, and Supheparin
refer to supernatants that had been in contact with the proteolytically
derived gelatin-binding domain, 120-kDa cell-binding domain, or
high-affinity heparin domain, respectively). After incubation, the
protein yields of all the supernatants, including
Sup0 and SupFN, were
approximately the same, indicating that the specific part of the
antibody directed to fibronectin constituted only a very small fraction
of the total amount of protein. To check whether adsorption on the
sepharoses was complete, we performed ELISAs in which we used intact
fibronectin, fibronectin fragments, or GST as first layer, the
supernatants as second layer, and peroxidase-conjugated swine
anti-rabbit antibodies as the third layer. SupFN
did not react with intact fibronectin or any of the fragments.
Sup0 and SupGST reacted
with intact fibronectin and all fragments, indicating that on each
fragment, epitopes were recognized by the polyclonal antibody. The
extent of interaction of Sup0 and
SupGST with intact fibronectin and the fragments
was the same, and no reaction was observed with GST. Supernatants that
had been in contact with fibronectin fragments coupled to Sepharose no
longer reacted with the same fragments in the ELISA, whereas the
reaction with nonoverlapping fragments had not changed compared with
Sup0/SupGST. The reaction
with partially overlapping fragments was partially reduced.
Perfusion Studies
Perfusion studies were performed in a parallel-plate perfusion
chamber with well-defined rheological characteristics designed to
accommodate duplicate glass coverslips.36 Whole blood
obtained by venipuncture from healthy volunteer donors was
anticoagulated with 1/10 volume 110 mmol/L trisodium citrate.
Whole blood (15 mL) was prewarmed at 37°C for 5 minutes and then
recirculated for 5 minutes at a wall shear rate of 300
s-1 through the perfusion chamber, which
contained 2 protein-sprayed coverslips or coverslips with ECM. In the
experiments in which we directly studied the effect of fragments FN4
and FN5 on adhesion, we added the fragments to the perfusate 10
minutes before the perfusion. After perfusion, the coverslips were
removed and rinsed with 10 mmol/L HEPES buffer containing 150
mmol/L NaCl (pH 7.35). They were then fixed with 0.5%
glutaraldehyde/PBS, dehydrated in methanol, and stained
with May-Grünwald-Giemsa stain as described
previously.36 Platelet adhesion was evaluated with a
light microscope at x1000 magnification, and the coverage was measured
with an Image Analyzer (AMS 40-10). Platelet coverage,
expressed as the percentage of the surface covered with platelets,
is the average of 60 fields per coverslip.
Statistical Analysis
Preincubation of ECM With Supernatants
The extent of inhibition by treatment with
Sup0 compared with treatment with
SupFN varied slightly among experiments
(36.7±5.8%, mean±SD; n=17), depending on the donor and the batch of
ECM that were used each day. Therefore, we compared the platelet
coverage found with a certain treatment to the values of minimal
(treatment with Sup0) and
maximal (treatment with SupFN) platelet
coverage by transforming values of platelet coverage on individual
coverslips according to the following equation:
 | (1) |
(0.05) was divided by
the total number of comparisons made,37 the so-called
Bonferroni correction, which resulted in a nominal significance level
of 0.0014.
Inhibition of Adhesion by FN4 and FN5
The absolute value of maximal coverage (no addition of fragment)
varied among experiments depending on the donor and the batch of ECM or
fibronectin that were used on each day. Therefore, we compared the
platelet coverage found with a certain concentration of fragment
FN4 or FN5 to this value by analyzing the ratio between them instead of
absolute differences in platelet coverage. For reasons of
normality, the ratios were logarithmically transformed. Two-way ANOVA
was performed with a sequential approach in which the effect of a
treatment was assessed after adjustment for the effect introduced by
different donors/different batches of ECM or fibronectin. To correct
for multiple comparisons, the significance level (0.05) was
divided by the number of comparisons made (fibronectin surface:
concentrations, =0.017; ECM: 4 concentrations, =0.0125). All
statistical analyses were performed according to Sokal and
Rohif.37
Results
Purification of Fibronectin Fragments
Proteolytic Fragments
Purified plasma fibronectin was digested for 3 hours with
cathepsin D, and the digest was applied to a gelatin-Sepharose and a
heparin-Sepharose column that were coupled in series. Figure 2 shows the 3 different fragments that we
eluted from the gelatin- or heparin-Sepharose column. The 70-kDa
gelatin-binding domain (lane 1), located at the amino-terminal end of
the molecule, eluted from the gelatin-Sepharose column with 6 mol/L
urea at low pH. The central 120-kDa cell-binding domain (lane 2) eluted
from the heparin-Sepharose column with 0.1 mol/L NaCl. A small fraction
of the central cell-binding domain was visible as a 110-kDa band. Both
the 120- and 110-kDa bands reacted on blot with an mAb directed to the
cell-attachment site on fibronectin (not shown). The high-affinity
heparin-binding domain (lane 3), located at the carboxy-terminal end of
the molecule, eluted from the heparin-Sepharose column with 0.5 mol/L
NaCl. As previously described, under reducing conditions, the
heparin-binding domain migrated as 2 separate bands of 60 and 70
kDa, because the type III connecting segment is present only in the
A chain of fibronectin.25 38 Both bands reacted on blot
with an mAb directed to the high-affinity heparin-binding site of
fibronectin (not shown).
|
Fusion Proteins
Fibronectin fragments FN1 to FN7 were expressed as fusion proteins
of GST. All proteins were purified from the soluble fraction of the
bacterial lysate. Figure 3A shows
SDS-PAGE analysis of the fusion proteins under reduced
conditions after affinity purification on glutathione-Sepharose beads.
All proteins migrated approximately at their expected molecular weight
(MW) based on their amino acid composition. With FN6 and FN6 RGE, and
to a lesser extent with FN1 and FN7, 2 or 3 lower bands with an MW of
26 kDa comigrated, which reacted on blot with a polyclonal antibody
directed to GST (not shown). Figure 3B shows that on blot, all
fusion proteins reacted with polyclonal anti-fibronectin antibodies. In
samples FN1, FN2, FN6, and FN6 RGE, some lower bands were visible,
which were probably degradation products of the fragments. Fragment
FN1 also showed a higher band, which may represent a dimer of
the protein. We performed densitometry on Coomassie brilliant
blue–stained gels to determine the proportion of intact fragment (ie,
fragment of expected MW) in each sample. We found a percentage of 98%
for FN3, FN4, and FN5; of 90% for FN1, FN2, and FN7; and of 50% for
FN6 and FN6 RGE.
|
Perfusion Studies
Adhesion Studies
First, we investigated the adhesive capacity of the fragments
themselves (Figure 4). Per coverslip, 100
µL of the indicated concentration was sprayed. Subsequently, they
were perfused for 5 minutes with whole blood at a shear rate of 300
s-1. Figure 4A shows adhesion to intact
fibronectin (FN), the 120-kDa cell-binding domain, and the 33- and
40-kDa fragments, whereas Figure 4B shows adhesion to FN4, FN5,
FN6, and GST alone as control for the fusion proteins. As previously
described,22 adhesion to intact fibronectin had its
optimum at 720 nmol/L (a spraying concentration of 720 nmol/L equals a
surface concentration of 5 µg/cm2), whereas a
higher concentration resulted in a lower platelet coverage because
of loss of platelet spreading. A similar pattern was observed for
the 120-kDa cell-binding domain, although the absolute amount of
coverage was higher than on intact fibronectin. Fragment FN6 started to
show adhesion at the same concentration as intact fibronectin and the
120-kDa fragment, but maximal adhesion was reached at a 2.5 times
higher concentration of 1.8 µmol/L. Furthermore, we did not
observe a decline in platelet coverage at a concentration of
3.6 µmol/L. For the other fragments as well, we did not observe
this decline in adhesion at higher concentrations. Fragment FN5 started
to show adhesion at a concentration of 0.7 µmol/L, reaching
maximal adhesion at 3.6 µmol/L. Fragment FN4 and the 33-kDa
fragment needed a minimal concentration of 1.8 µmol/L, whereas
maximal adhesion was reached at 7.3 µmol/L. The 40-kDa fragment
was an exception compared with the other proteins, because it showed a
coverage of only 8% over the whole range between 0.9 and 7.3
µmol/L. All fragments supported both initial adhesion and
platelet spreading. On the basis of the concentration that resulted
in half-maximal adhesion (EC50), we can place the
fragments in the following order of activity: intact
fibronectin/120-kDa fragment (0.3 µmol/L) > FN6 (0.55
µmol/L) > FN5 (1.5 µmol/L) > FN4 (2.5
µmol/L) >33-kDa fragment (3.5 µmol/L) (we left out the 40-kDa
fragment because of its aberrant behavior). The GST protein itself did
not support adhesion (Figure 4B), nor did the fusion proteins
FN1, FN2, FN3, FN7, and FN6 RGE (not shown). The proteolytically
derived gelatin- and heparin-binding domains also were unable to
support adhesion (not shown).
|
Inhibition Experiments
Preincubation of ECM With Different Supernatants
To investigate the role of different domains of fibronectin in
adhesion to fibronectin in a more complex system, we circulated blood
over ECM. We performed our experiments at a shear rate of 300
s-1, where adhesion to the ECM had previously
been described to depend on fibronectin in the matrix but not on the
RGD site in this molecule.23
Polyclonal anti-fibronectin antibodies were incubated with Sepharose to
which intact fibronectin, no protein, or a fibronectin fragment had
been coupled. In this way, we obtained supernatants in which a specific
part of the antibodies had been removed. Treatment of the ECM with
SupFN from which all anti-fibronectin antibodies
were removed and Sup0 that still contained all
anti-fibronectin antibodies was considered to result in maximal and
minimal platelet coverage, respectively. Treatment with
Sup0 showed a dose-dependent inhibition of
adhesion. At 750 µg/mL, the inhibition was 32.2±1.5%, whereas a
concentration of 1 mg/mL resulted in an inhibition of 34.3±3.5%
(mean±SEM; n=3), indicating that 750 µg/mL was sufficient to reach
maximal inhibition (Figure 5). For the
subsequent experiments (Table 2), we
preincubated the ECM with a concentration of 1 mg/mL of the
supernatants. Treatment with supernatants that had been in contact with
either the gelatin-binding domain (Supgelatin) or
the heparin-binding domain (Supheparin and
SupFN7) showed no increase of coverage compared
with treatment with Sup0. These data suggested
that these segments of fibronectin do not contain adhesive sequences
for platelets and that all the information needed for platelets
to adhere resides in the 120-kDa central cell-binding domain. Indeed,
treatment with Supcell resulted in a complete
restoration of adhesion to the level of treatment with
SupFN. To study in detail the role of different
type III repeats in this region, we prepared a series of recombinant
fragments of fibronectin as fusion proteins with GST. Treatment with
SupGST as a control for the fusion proteins
showed no increase of coverage compared with treatment with
Sup0, ensuring that a possible positive effect on
adhesion by incubation with other supernatants would not be caused by
the GST part of the protein. Treatment with supernatants covering III-1
up to III-7 (SupFN1 and
SupFN2) or covering only III-10
(SupFN4) showed no increase of coverage compared
with treatment with Sup0. Treatment with
Sup40kDa showed some increase of adhesion
compared with treatment with Sup0, but this
difference was not statistically significant. Treatment with a series
of other supernatants, SupFN3,
SupFN5, SupFN6,
SupFN6 RGE, or Sup33kDa,
led to partial restoration of adhesion: the coverage found was
significantly (P<0.0005) different from the coverage on
coverslips incubated with either Sup0 or
SupFN. The extent of restoration of adhesion was
slightly different among these supernatants (Figure 6). Treatment with
SupFN6 showed the greatest effect, with a 63%
increase in adhesion from the value of coverage found with
Sup0 to the value of coverage observed with
SupFN. For SupFN3,
SupFN5, SupFN6 RGE, and
Sup33kDa, the increase of adhesion was 48%,
54%, 44%, and 50%, respectively. However, on the basis of
experiments in which the effects of both supernatants were tested
simultaneously, we found no significant difference in
effect between SupFN6 and
Sup33, SupFN6 and
SupFN5, SupFN6 and
SupFN6 RGE, and SupFN3 and
SupFN5, respectively.
|
|
|
Addition of Fragment FN4 or FN5 to the Perfusate
In the previous experiments, we found that preincubation of the
ECM with SupFN4, as with
Sup0, resulted in maximal inhibition of adhesion,
whereas preincubation with SupFN5 showed partial
recovery of adhesion compared with preincubation with
Sup0. This suggested that domain III-9 of
fibronectin contains adhesive sequence(s). In the following
experiments, we tested the direct effect of these 2 fragments on
adhesion. Both FN4 and FN5 showed a dose-dependent inhibition of
adhesion to isolated fibronectin, reaching a maximal inhibition of
54.4% and 85.4%, respectively, at a concentration of 25 µmol/L
(Figure 7A). ANOVA in which we
tested the effect of different concentrations per fragment showed
significant (P<0.016) reduction of adhesion for fragment
FN5 at all concentrations and for fragment FN4 at 10 and 25
µmol/L. The IC50 for FN4 and FN5 were 5.7 and
4 µmol/L, respectively. Then we investigated the effect of the
fragments on adhesion to the ECM (Figure 7B). Fragment FN4
showed no decrease in coverage over the whole range of concentrations
up to a concentration of 100 µmol/L. Fragment FN5 significantly
(P<0.0125) reduced adhesion at 50 and 100 µmol/L.
The IC50 was 13.7 µmol/L.
|
) or FN5 (•) were added. The
preincubation time of the fragments was 15 minutes. Platelet
coverage is expressed as percentage of the coverage found when no
fragment was added. Values represent mean±SEM of 4
(fibronectin surface) or 5 (ECM) separate experiments, each performed
in quadruplicate. Statistical analysis was performed as
described under Experimental Procedures. *P<0.005;
**P<0.01.Discussion
In the present work, we investigated the involvement of different fibronectin domains in fibronectin-dependent platelet adhesion to the ECM, a system that had previously been shown to be a suitable model for the subendothelium.33 To do so, we prepared a series of fragments, including both proteolytically derived and recombinant fragments. Observations with respect to adhesion to the ECM were compared with results that we obtained in experiments in which we used the fragments themselves as adhesive surface.
Platelet adhesion to isolated fibronectin fully depended on the 120-kDa central cell-binding domain of the molecule: (1) the 120-kDa cell-binding domain and the total fibronectin molecule were equally active (both IC50 of 0.3 µmol/L), and (2) fragments covering the gelatin- or heparin-binding domain were not able to support adhesion. Examination of the adhesive capacity of fragments covering this central region showed that for platelet adhesion to the isolated protein, the RGD sequence was absolutely required. Changing this sequence into RGE, as we did with fragment FN6 RGE, resulted in complete loss of adhesion. All other fragments that did not contain this site, namely FN1, FN2, and FN3, were also unable to support adhesion. The fragments that did support adhesion showed slight differences in activity. The fusion protein FN6 was almost as active as the cell-binding domain (IC50 of 0.55 µmol/L), whereas FN5, FN4, and the 33-kDa fragment showed a slightly increasing IC50 of 1.5, 2.5, and 3.5 µmol/L, respectively. The difference in adhesive capacity between intact fibronectin, the 120-kDa cell-binding domain, and FN6 on one hand, and FN5, FN4, and the 33-kDa fragment on the other hand, may reflect differences in conformation or may be caused by the presence of additional adhesive sites in the larger proteins. A specific conformation of the 40-kDa fragment, a fragment larger than fragment FN4, which supported adhesion very well, may account for the relative inability of this fragment to support adhesion. This may be a consequence of inappropriate folding because the start and end points of the 40-kDa fragment do not correspond to repeat boundaries. It may also be a consequence of the attachment of the protein to the glass coverslip. Previous studies showed that the CS1 region of fibronectin lost activity on adsorption to plastic substrate,39 which was regained by conjugating the peptide to a carrier protein.40 Similarly, attachment of the 40-kDa fragment, which was not a fusion protein and thus without carrier,30 may have resulted in a conformation in which adhesive site(s) were shielded.
With respect to the activity of the different fragments, we found both similarities and differences between platelet adhesion to isolated fibronectin, as described above, and fibronectin in the ECM. As with adhesion to the isolated protein, we found that fibronectin-dependent adhesion to the ECM also completely depended on the 120-kDa central cell-binding domain: (1) preincubation of the polyclonal antibody with the gelatin- or heparin-binding domain did not result in a decrease of the inhibitory capacity of the antibody, and (2) preincubation of the antibody with the 120-kDa cell-binding domain resulted in a complete loss of inhibition of adhesion equal to the loss of inhibition found when the antibody was preincubated with intact fibronectin. Further analysis of this central region showed that adhesion to fibronectin in the ECM, in contrast to adhesion to isolated fibronectin, depended on III-9 instead of III-10: a series of supernatants, which had in common that they were all preincubated with a fragment that contained a complete III-9 (SupFN6, SupFN5, Sup33, SupFN3, and SupFN6 RGE), displayed a partial increase (50%) in coverage. In contrast, treatment with supernatants preincubated with fragments that did not comprise III-9 (SupFN1, SupFN2, and SupFN4) or contained an incomplete III-9 (Sup40kDa) did not lead to increased coverage. We found further that direct addition of a fragment consisting of only III-10 (FN4) did not decrease the adhesion to the ECM, whereas a fragment containing both III-9 and III-10 (FN5) did reduce adhesion. These data suggest that III-9 contains specific adhesive sequences for platelets adhering to the ECM, whereas III-10, in which the RGD sequence is located, apparently does not contribute to adhesion. To check whether the polyclonal anti-fibronectin antibodies contained inhibitory antibodies directed to III-10 at all, we performed control perfusions in which we used fragment FN4 itself as adhesive surface. Adhesion to this fragment was completely inhibited by the polyclonal antibodies (not shown), indicating that indeed, inhibitory antibodies directed to this repeat were present. These antibodies presumably inhibited the interaction between platelets and the RGD site, because we had found, as described above, that platelet adhesion to fragment FN6 fully depended on this site. Unfortunately, no MAbs directed against FN are known that inhibit adhesion to ECM, and therefore, we must rely on the preadsorbed polyclonal antibodies.
In this article, we present evidence that III-9 supports platelet adhesion to the ECM. However, we cannot completely exclude the possibility that other regions also contribute, because we did not reach complete restoration of adhesion with our fusion proteins, as we did with the 120-kDa cell-binding domain, which was derived from purified plasma fibronectin. We cannot explain this partial restoration, especially because we found that the inhibition caused by direct addition of 100 µmol/L of fragment FN5 to the perfusate (34.0±7.0%, mean±SD; n=5) equaled the inhibition that we observed when we preincubated the ECM with Sup0 (36.7±5.8%; n=17). A reason might be incomplete absorption of the polyclonal antibodies by the different Sepharoses. However, as described under Experimental Procedures, we established with ELISAs that absorption was complete, indicating that this is not a likely explanation and leaving the question of partial restoration yet unanswered.
Different adhesive sequences have been mapped to III-9. Using chimeric
proteins consisting of III-10 and a III-9 in which different amino
acids were substituted for the corresponding amino acids in III-8, Aota
et al19 identified a stretch of 5 amino acids,
Pro1376-Asn1380, that
could competitively inhibit VLA-5–mediated cell spreading on
fibronectin and could support this cell spreading itself. In the same
article, they deduced another sequence,
Asp1373-Pro1375, which they
assumed to contain adhesive activity as well and which could operate
independently from sequence
Pro1376-Asn1380.
Bowditch et al21 described almost the same sequence
(Asp1373-Thr1313) as
recognition motif for purified GP IIb/IIIa and as inhibitor
of ADP-stimulated platelet aggregation. Although both VLA-5 and GP
IIb/IIIa appeared to recognize the same sequence, the importance of the
respective amino acids in this sequence was different for the 2
receptors. It is conceivable that the sequence
Asp1373-Thr1383 also plays
a role in fibronectin-dependent adhesion to the ECM, especially because
treatment with Sup40kDa did not restore adhesion
and the 40-kDa fragment starts at amino acid
Asn1380. However, several lines of evidence argue
against this. (1) As stated above, the sequence is recognized by VLA-5
and GP IIb/IIIa. However, we found that antibodies directed to these
receptors did not inhibit adhesion to the ECM.22 (2)
Bowditch et al21 showed that binding of III-10 to GP
IIb/IIIa was inhibited by this sequence, suggesting that
Asp1371-Thr1383 and the RGD
sequence represent 2 mutually exclusive sites on fibronectin
that compete for binding to this receptor. Such a phenomenon has been
described for the RGD site and 
400 to 411 peptide in
fibrinogen.41 Previous studies, however, showed that
platelet adhesion to the ECM was not inhibited by RGD-containing
peptides.23 Thus, this sequence is not likely to support
platelet adhesion to the ECM. Another candidate may be the
amino-terminal part of III-9; Katayama et al16 described
an antibody (FN30-8), which inhibited cell adhesion to
fibronectin and recognized a fragment of 76 amino acids, which
contained part of III-8 and the first 14 amino acids of III-9. If this
antibody turns out to recognize this stretch of amino acids in III-9,
an adhesive sequence for platelet may be located in this
region.
At present, we do not know which platelet receptor will recognize III-9. Tanabe et al42 described a 29-kDa fragment of fibronectin, spanning the high-affinity heparin domain in which the RGD site was not contained, which bound to purified GP IIb/IIIa and to thrombin-stimulated platelets. This latter interaction was not inhibited by an RGD-containing peptide and an anti–GP IIb/IIIa antibody. This example shows us that, although we previously established that an RGD-containing peptide or an antibody directed to GP IIb/IIIa or VLA-5 did not inhibit adhesion to the ECM,22 23 this does not necessarily mean that GP IIb/IIIa or VLA-5 is not involved in this process. We can only speculate why the RGD sequence in matrix fibronectin is not available for platelet adhesion. There are conformational differences between matrix fibronectin and soluble fibronectin coated on glass. Soluble fibronectin is a dimeric glycoprotein with no tendency to polymerize. Matrix fibronectin is an insoluble multimeric form with numerous additional interactions with itself and with other matrix components. There are indications that the RDG-containing III-10 module of fibronectin is involved in matrix assembly.43 Interaction between the III-10 module with the 70-kD amino-terminal fragment appears to be necessary to allow assembly of fibronectin via this 70-kD fragment into the matrix. Because of this interaction, it is possible that the RDG sequence in III-10 is no longer accessible for integrin binding.
In conclusion, our results suggest that repeat III-9 of fibronectin supports platelet adhesion to the ECM independently from II-10 and, as a consequence, from the RGD site here. In this respect, the adhesion to the ECM differed from adhesion to the isolated protein for which, on the basis of the comparison of the adhesive capacity of fragments FN6 and FN6 RGE, we assume the RGD site to be crucial. Further research will be necessary to identify the exact sequence in III-9 that is recognized and the nature of the receptor involved.
Acknowledgments
This study was supported by the Dutch Heart Foundation (grant 88219) and the foundation De Drie Lichten in the Netherlands. We gratefully acknowledge Deane Mosher (University of Wisconsin, Madison) for providing us with the plasmid pFH100. We thank A. van den Hoeven for the culturing of the endothelial cells. We also thank the Department of Enzymology and Protein Engineering, Utrecht University, Netherlands, for giving us the opportunity to culture large batches of bacteria. The Red Cross Bloodbank Utrecht (Utrecht, Netherlands) is acknowledged for supplying large quantities of fresh blood. We further thank L. van der Tweel (Center of Biostatistics, Utrecht University, Netherlands) for helpful advice on the statistical analyses and Dr G.H. van Zanten (Department of Hematology, University Hospital Utrecht, Netherlands) for critical advice on the manuscript.
Received January 11, 1999; accepted November 4, 1999.
References
1. Hynes RO. Fibronectins. New York, NY: Springer-Verlag; 1990.
2.
Petersen TE, Thogersen HC, Skorstengaard K,
Vibe-Pedersen K, Sahi P, Sottrup-Jensen L, Magnusson S. Partial primary
structure of bovine plasma fibronectin: three types of internal
homology. Proc Natl Acad Sci U S A. 1983;80:137–141.
3. Kornblihtt AR, Umezawa K, Vibe-Pedersen K, Baralle FE. Primary structure of human fibronectin: differential splicing may generate at least 10 polypeptides from a single gene. EMBO J.. 1985;4:1755–1759.[Medline] [Order article via Infotrieve]
4. Pierschbacher MD, Ruoslahti E. Cell attachment activity of fibronectin can be duplicated by small synthetic fragments of the molecule. Nature. 1984;309:30–33.[Medline] [Order article via Infotrieve]
5.
Pytela R, Pierschbacher MD, Ginsberg MH, Plow EF,
Ruoslahti E. Platelet membrane glycoprotein IIb/IIIa:
member of a family of Arg-Gly-Asp-specific adhesion receptors.
Science. 1986;231:1559–1562.
6.
Parise LV, Phillips DR. Fibronectin-binding properties
of the purified platelet glycoprotein IIb-IIIa complex.
J Biol Chem. 1986;261:14011–14017.
7.
Hautanen A, Gailit J, Mann DM, Ruoslahti E. Effects of
modifications of the RGD sequence and its context on recognition by the
fibronectin receptor. J Biol Chem. 1989;264:1437–1442.
8.
Giancotti FG, Languino LR, Zanetti A, Peri G, Tarone
G, Dejana E. Platelets express a membrane complex immunologically
related to the fibroblast fibronectin receptor and distinct from GP
IIb/IIIa. Blood. 1987;69:1535–1538.
9.
Piotrowicz RS, Orchekowski RP, Nugent DJ, Yamada KY,
Kunicki TJ. Glycoprotein Ic-IIa functions as an
activation-independent fibronectin receptor on human platelets.
J Cell Biol. 1988;106:1359–1364.
10.
Wayner EA, Carter WG, Piotrowicz RS, Kunicki TJ. The
function of multiple extracellular matrix receptors in mediating cell
adhesion to extracellular matrix: preparation of monoclonal antibodies
to the fibronectin receptor that specifically inhibit cell adhesion to
fibronectin and react with platelet glycoproteins
Ic-IIa. J Cell Biol. 1988;107:1881–1891.
11. Hynes RO. lntegrins: versatility, modulation, and signaling in cell adhesion. Cell. 1992;69:11–25.[Medline] [Order article via Infotrieve]
12.
Akiyama SK, Yamada KM. Synthetic peptides competitively
inhibit both direct binding to fibroblasts and functional biological
assays for the purified cell-binding domain of fibronectin.
J Biol Chem. 1985;260:10402–10404.
13.
Akiyama SK, Hasegawa E, Hasegawa T, Yamada KM. The
interaction of fibronectin with fibroblastic cells. J Biol
Chem. 1985;260:13256–13260.
14.
Streeter HB, Rees DA. Fibroblast adhesion to RGDS shows
novel features compared with fibronectin. J Cell Biol. 1987;105:507–515.
15. Obara M, Kang MS, Yamada KM. Site-directed mutagenesis of the cell-binding domain of human fibronectin: separable, synergistic sites mediate adhesive function. Cell. 1988;53:649–657.[Medline] [Order article via Infotrieve]
16. Katayama M, Hino F, Odate Y, Goto S, Kimizuka F, Kato I, Titani K, Sekiguchi K. Isolation and characterization of two monoclonal antibodies that recognize remote epitopes on the cell-binding domain of human fibronectin. Exp Cell Res. 1989;185:229–236.[Medline] [Order article via Infotrieve]
17.
Aota S, Nagai T, Yamada KM. Characterization of regions
of fibronectin besides the arginine-glycine-aspartic acid sequence
required for adhesive function of the cell-binding domain using
site-directed mutagenesis. J Biol Chem. 1991;266:15938–15943.
18.
Nagai T, Yamakawa N, Aota S, Yamada SS, Akiyama SK,
Olden K, Yamada M. Monoclonal antibody characterization of two distant
sites required for function of the central cell-binding domain of
fibronectin in cell adhesion, cell migration, and matrix assembly.
J Cell Biol. 1991;114:1295–1305.
19.
Aota S, Nomizu M, Yamada KM. The short amino acid
sequence Pro-His-Ser-Arg-Asn in human fibronectin enhances
cell-adhesive function. J Biol Chem. 1994;269:24756–24761.
20.
Bowditch RD, Halloran CE, Aota S, Obara M, Plow EF,
Yamada KM, Ginsberg MH. Integrin
IIbß3 (platelet
GPIIb-IIIa) recognizes multiple sites in fibronectin. J Biol
Chem. 1991;266:23323–23328.
21.
Bowditch RD, Hariharan M, Tominna EF, Smith JW, Yamada
KM, Getzoff ED, Ginsberg MH. Identification of a novel integrin binding
site in fibronectin. J Biol Chem. 1994;269:10856–10863.
22.
Beumer S, IJsseldijk MJW, de Groot PG, Sixma JJ.
Platelet adhesion to fibronectin in flow: dependence on surface
concentration and shear rate, role of membrane
glycoproteins GP IIb/IIIa and VLA-5, and inhibition by
heparin. Blood. 1994;84:3724–3733.
23.
Nievelstein PFEM, Sixma JJ. Glycoprotein
IIb-IIIa and RGD(S) are not important for fibronectin dependent
adhesion under flow conditions. Blood. 1988;72:82–88.
24. Kiebe RJ, Bentley KL, Sasser PJ, Schoen RC. Elution of fibronectin from collagen with chaotrophic agents. Exp Cell Res. 1980;130:111–117.[Medline] [Order article via Infotrieve]
25. Richter H, Seidi M, Hdrmann H. Location of heparin-binding sites of fibronectin: detection of a hitherto unrecognized transamidase sensitive site. Hoppe Seylers Z Physiol Chem. 1981;362:399–408.[Medline] [Order article via Infotrieve]
26. Smith DB, Johnson KS. Single-step purification of polypeptides expressed in Escherichia coli as fusion with glutathione-S-transferase. Gene. 1988;67:31–40.[Medline] [Order article via Infotrieve]
27. Dufour S, Gutman A, Bois F, Lamb N, Thierry JP, Kornblihtt AR. Generation of full-length cDNA recombinant vectors for the transient expression of human fibronectin in mammalian cell lines. Exp Cell Res. 1991;193:331–338.[Medline] [Order article via Infotrieve]
28.
Yang Y-S, Watson WJ, Tucker PW, Capra JD. Construction
of recombinant DNA by exonuclease recession. Nucleic Acids
Res. 1993;21:1889–1893.
29.
Sanger F, Nicklen S, Coulson AR. DNA sequencing with
chain termination inhibitors. Proc Natl Acad Sci
U S A. 1977;74:5463–5467.
30. Vogel T, Werber MM, Guy R, Levanon A, Nimrod A, Legrand C, Gorecki M, Eldor A, Panet A. Studies on fibronectin and its domains, I: novel recombinant cell-binding domain of fibronectin: a modulator of human platelet functions. Arch Biochem Biophys. 1993;300:501–509.[Medline] [Order article via Infotrieve]
31. De Ronde A, Stam JG, Boers P, Langedijk H, Meioen R, Hesselink W, Keldermans LCEJM, van Vliet A, Vershoor EJ, Horzinek MC, Egberink HF. Antibody response in cats to the envelope proteins of feline immunodeficiency virus: identification of an immunodominant neutralization domain. Virology. 1994;198:257–264.[Medline] [Order article via Infotrieve]
32. Laemmli UK. Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Nature. 1970;227:680–685.[Medline] [Order article via Infotrieve]
33. Sixma JJ, Nievelstein PFEM, Zwaginga J, de Groot PG. Adhesion of blood platelets to the extracellular matrix of cultured endothelial cells. Ann N Y Acad Sci. 1987;516:39–51.[Medline] [Order article via Infotrieve]
34. Wu YP, Sixma JJ, de Groot PG. Cultured endothelial cells regulate platelet adhesion to their extracellular matrix by regulating its von Willebrand content. Thromb Haemost. 1995;73:713–718.[Medline] [Order article via Infotrieve]
35.
Aznar-Salatti J, Bastida E, Haas TA, Escolar G, Ordinas
A, de Groot PG, Buchanan MR. Platelet adhesion to exposed
endothelial cell extracellular matrixes is influenced
by the method of preparation. Arterioscler Thromb. 1991;11:436–442.
36. Sakariassen KS, Aarts PAMM, de Groot PG, Houdijk WPM, Sixma JJ. A perfusion chamber developed to investigate platelet interaction in flowing blood with human vessel wall cells, their extracellular matrix, and purified components. J Lab Clin Med. 1983;102:522–535.[Medline] [Order article via Infotrieve]
37. Sokal RR, Rohif FJ. Biometry: The Principles and Practice of Statistics in Biological Research. San Francisco, Calif: WH Truman & Co; 1962.
38. Tressel T, McCarthy JB, Calaycay J, Lee TD, Legesse K, Shively JE, Pande H. Human plasma fibronectin. Biochem J.. 1991;274:731–738.
39.
Humphries MJ, Akiyama SK, Komoriya K, Olden K, Yamada
KM. Identification of an alternative spliced site in human fibronectin
that mediates cell-type specific adhesion. J Cell Biol. 1986;103:2637–2647.
40.
Humphries MJ, Komoriya A, Akiyama SK, Olden K, Yamada
KM. Identification of two distinct regions of the type III connecting
segment of human plasma fibronectin that promote cell type-specific
adhesion. J Biol Chem. 1987;262:6886–6892.
41.
Lam SC-T, Plow EF, Smith MA, Andrieux A, Ryckwaert J-J,
Margerie G, Ginsberg MH. Evidence that arginyl-glycyl-aspartate
peptides and fibrinogen gamma chain peptides share a common binding
site on platelets. J Biol Chem. 1987;262:947–950.
42.
Tanabe J, Fujita H, Iwamatsu A, Mohri H, Ohkubo T.
Fibronectin inhibits platelet aggregation independently of RGD
sequence. J Biol Chem. 1993;268:27143–27147.
43.
Magnussen MK, Mosker DF. Fibronectin, structure,
assembly and cardiovascular implications. Thromb
Vasc Biol. 1998;18:1363–1370.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||